# PEP? Are Those Power Meters Showing Us PEP Peak Envelope Power, or just Peak Power?

Discussion in 'Amateur Radio Amplifiers' started by KQ4X, Dec 31, 2012.

Not open for further replies.

PEP or Average power should always be based on RMS voltages. Peak voltage is used by the Hi Fi crowd to inflate amplifier ratings. I suspect RM Italy falls into the same deceptive advertising, since all of their amplifier PEP ratings are based on 2X maximum saturated power for the devices used in their SS amps.

Pete

2. ### AA5CTHam MemberQRZ Page

G0HZU, I have a question regarding the derivation of the equation above, in particular, where did the value "100" come from?

Jim WB5WPA

3. ### K7JEMHam MemberQRZ Page

I think people are trying to read more into this than actually exists. The first thing we need to do is go back and look at the definition of what PEP is:

"Peak envelope power (PEP) is the average power supplied to the antenna transmission line by a transmitter during one radio frequency cycle at the crest of the modulation envelope, under normal operating conditions."

If we look at that, we see that PEP is an average power, just like the average power we see for FM or a key down CW transmission. The only difference is that FM is averaged over a million RF cycles (or more), while PEP is averaged over a single RF cycle.

Now, when we look at the rest of the definition, it refers to normal operating conditions. For a SSB TX, that would mean that the highest modulation frequency would be around 3KHz or less, making the peak or the crest somewhat around 100uS. That would be the timeframe of the peak, or crest of modulation, the point where the average would be taken. But we are not looking at a ragged low frequency modulation signal, we are looking at near pure sine waves which comprise the RF portion of the signal. At 1MHz, the number of RF cycles that are contained in that 100uS crest is equal to 100. At higher frequencies, like we use on HF, the number may be many times that amount.

So, no complex math or understanding is needed, since by definition, we are looking at a single sine wave lasting less than a uS in duration. Most meters would be really incapable of seeing this, but since the crest lasts 100uS, or more, a meter could be designed to capture and hold that peak reading, and give us a fairly accurate PEP readout. Certainly a digital circuit could be built to do this, by measuring and holding the highest reading. The analog way to do this is by charging a capacitor quickly, and having a long time constant decay.

The ham grade PEP meters are probably fairly accurate if they have electronics in them to quickly charge a capacitor, and a buffer amp to allow a very slow decay. As Steve mentioned earlier, a good way to test one is to transmit a single "dit" on CW, and see if that PEP reading matches the key down power of the CW transmitter. If it does, then the meter will be fairly accurate for SSB readings. If it doesn't, then it will not be accurate on SSB either.

Joe

4. ### G0HZUQRZ Member

It looks a bit odd at first but it's just a shorthand way of doing the sum without having to convert the Vpeak to Vrms.

for a sine wave, (Vpeak*Vpeak)/2 = (Vrms *Vrms) so you can exploit this to simplify the power calculation. So the 50 ohms gets multiplied by 2 so that's where the "100" comes from.

Here's a 50 ohm system example:

if Vrms = 1V and Vpk = 1.414V (for a sine wave) then the regular calculation gives (1*1)/50 for the power but if you measured Vpk then the quick way to the same answer is to do (1.414*1.414)/100. Both should give 0.02W for a 50 ohm system.

5. ### NZ9YHam MemberQRZ Page

I hear guys often lightly accuse others of running over limit power. They dont seem to understand the huge db gains over thier own stations would require 10s or hundreds times legal limit. If those complainers only knew that antenna are where the real gains come from........

6. ### WB2WIKPlatinum SubscriberPlatinum SubscriberQRZ Page

Antennas, location and time-specific propagation are the major players. If I want to be 20 dB stronger over a particular path, I can wait until prop peaks for the path and pick up 20 dB -- sometimes only for a few minutes.

Because that window can be so brief, it might be good only for "me" and almost nobody else unless they're my neighbors. Everybody else will see the peak at a different time, possibly minutes different, or could be a lot longer.

Strongest signals on the "higher" bands are often from stations with advantageous locations, too. There are some guys in Europe who live on cliffs overlooking the sea, with a completely clear shot over the ocean from there to NA. They're all stronger than the average stations are.

7. ### WD4OOZHam MemberQRZ Page

Wow...what a discussion....No one has mentioned the easiest way to determine peak power for any type of modulation...cw, ssb, swept tone, etc. There are devices called TRUE RMS TO DC CONVERTERS. This will convert ANY signal to TRUE RMS values that now make it easy to calibrate a meter for PEAK or RMS, or what ever you want. They now set a standard value for any type of modulated power. Most of the Current Power meters that we have are cheap means of Peak power measurement and the Power readings measured will vary with duty cycle. Some will show the deferrence between a dit and a dah power levels and with SSB voice levels the Duty Cycle is totally variable and the meter is calibrate to be close. The added Cap network to get the peak reading is only good for the Duty Cycle of the Calibrating modulated waveform and all others will read differently.
True RMS to DC Converters are not that expensive but they do increase the cost of a product in a very conpetitive business.
Good Day and Good DX

WD4OOZ

Last edited: Jan 7, 2013
8. ### KH6AQHam MemberQRZ Page

Joe, you have explained it perfectly.

9. ### G0HZUQRZ Member

Yes this is a good thread

I do think 'some' basic analysis and maths is needed in order to understand the limitations of a typical meter that uses a peak voltage detector.
This limitation affects both PEP readings and also average power readings for cw signals.

For example if there are fairly significant harmonic signals present then the classic peak detector method used in many ham meters starts to
suffer from high levels of measurement uncertainty for measuring the power of a transmitter. Much higher levels of uncertainty compared to other
types of power meter. This is because you cannot estimate the true rms voltage of a distorted sine wave by just measuring the peak voltage of the
distorted waveform and doing a basic sum or scaling correction on the meter.

eg for a power meter using a peak voltage detector the measurement uncertainty caused by a harmonic at a certain -dBc level can be approximated as below:

-46dBc Uncertainty = approx +/-1%
-40dBc Uncertainty = approx +/-2%
-30dBc Uncertainty = approx +/-6%
-25dBc Uncertainty = approx +/-11%
-20dBc Uncertainty = approx +/-20%

The meter reading of this type of power meter is affected by the phase of the harmonic relative to the fundamental and this gives rise to the big
uncertainty window. eg the harmonic can make the indicated power appear larger or smaller depending on the relative phase. Hence the large
uncertainty figures above.

If you could rotate the phase of a -30dBc harmonic you would see the meter needle move back and forth by about +/-6% in power indication.
if the meter itself also has a quoted accuracy of +/- 5% for a pure sinewave then you can see the presence of harmonics at maybe -30dBc can
degrade the accuracy of this type of power meter a lot. That's why Bird recommend you keep harmonics at -50dBc for some of their power meters
that use peak voltage detectors. eg meters with peak detector diodes operating in the linear region.

Now you can argue that the typical limits for harmonics are -40dBc on a ham transmitter so maybe this has limited relevance, but if you want to
measure the average power from a homemade solid state pushpull amplifier that hasn't yet got a suitable LPF on the output, or you want to
measure the power from an unfiltered driver stage then the measurement uncertainty of the Vpeak detector power meter will be very poor
compared to a meter that can measure the average power correctly in the presence of harmonics.

So I think it is worth analysing this type of meter in more detail in order to understand the potential measurement uncertainty of a simple metering
system that is based on measuring just peak RF voltage (rather than true rms voltage).

10. ### KQ4XHam MemberQRZ Page

I have a novel idea (since I'm the one who started this discussion): Let's get the FCC to agree to go back to measuring our power limit by using INPUT POWER to the final amp. We could limit input power to 2,100 watts. Example would be: 3,000 volts @ .700 amps = 2,100 watts INPUT; and at even a high efficiency rate of 75% output, we'd be putting out 1,575 watts.

And it wouldn't be too difficult to watch our amp's voltmeter and "plate" amp meter in SSB mode to make sure we're not INPUTTING more than the 2,100 watts (or whatever the limit would be set at).

At least we could be free from the issues/problems at hand. I honestly think that would serve us all better. Does anyone agree with that?